Announcements => Announcements => Topic started by: Michael on August 29, 2008, 02:08:00 PM

Title: Rootvg datacenter is moving!
Post by: Michael on August 29, 2008, 02:08:00 PM
All rather sudden, so please accept our apologies for any downtime.
Title: Re: Rootvg datacenter is moving!
Post by: John R Peck on August 30, 2008, 11:48:30 AM

Well here we are again  ;D

This server, which is a multi-service pSeries machine, has now moved physical location and is back
up and running almost as before. 

Just one more technical issue is to be sorted out concerning the difference between dedicated IP
and name-based virtual hosting.  This caused longer than expected outage problems on the service,
something around 9 hours on Friday 29th August from 15:30.  It took a while to notice there was
any problem and implement a workaround - other sites on this server had only a 3 hour outage
while the DNS switched and the server was in physical transit. 

Three weeks ago we were informed of a massive increase in future co-location hosting and
traffic fees from our ISP, where the server had been housed in a shared data centre for the
last eight years.  We were forced to build our own dedicated server hosting room and
there was doubt it would be ready in time to take the server.  At the eleventh hour before
the mega-ransom-bill kicked in, we were able to get the connections in place ready to move.
So again apologies for the short notice of this change and outage.

Presumably Michael will come up with some new htaccess re-write rules or changes to the
site page coding to allow the site to revert to the planned name-based virtual hosting
without having to have it as the default IP-based host section where a multi-site switch
menu was intended, or I may just implement a series of links to allow the files to be found
from the unintended directory. 

If you have any further problems with the service, please do contact us with details.

Title: Re: Rootvg datacenter is moving!
Post by: John R Peck on September 01, 2008, 08:42:04 AM

And woopsie, another little web-only outage over Sunday 31st night after I messed up an httpd.conf
edit which kicked in (or rather didn't) after a UPS test failed - yet another component cooked to
death in the last data centre I suspect, or call it old age - anyway UPS replaced immediately,
everything has now been tested to destruction  ::)